Request for Performance Results on LongBench with llama3.1-8b-instruct-128k

[xlim1997]

Hi,

Could you please provide the results you obtained for llama3.1-8b-instruct-128k model under the ds method and the full attention method on the LongBench datasets?

[andy-yang-1]

@xlim1996 Hi, you can directly refer to LongBench/longbench_ds.sh